Apple integrates ML denoising into RAW 9 engine for OS 27
Apple's new RAW 9 engine, shipping with OS 27, integrates machine-learning denoising directly into Core Image at the demosaicing stage. This update promises sharper results across Apple hardware and 784 supported cameras, impacting content creation workflows for video professionals using Apple's ecosystem.
Key Takeaways
- RAW 9 integrates noise reduction into the demosaicing step rather than applying it as a post-process pass.
- System-level support extends to 784 camera models from vendors including Sony, Canon, and Fujifilm.
- The updated Core Image engine will be available this fall on iOS 27, iPadOS 27, visionOS 27, and macOS 27.
- Third-party developers must manually opt in to RAW 9 due to its higher processing demands compared to RAW 8.
Why It Matters
By moving ML denoising to the core demosaicing stage, Apple is commoditizing high-end image restoration that previously required expensive third-party plugins. This shift specifically benefits mobile-first video creators and developers of indie editing apps who can now access sophisticated noise reduction via the CIRAWFilter API without building proprietary models. For the broader ecosystem, it signals a move toward silicon-optimized, on-device AI that leverages the Neural Engine to outpace cloud-based or generic software solutions. Watch for Adobe and other pro-tool vendors to see if they maintain proprietary engines or eventually yield to Apple’s system-level performance gains on Apple Silicon.
Additional Context
The introduction of RAW 9 coincides with a broader push for 'Siri AI' and 'Apple Intelligence' across the 2026 software lineup. Per AppleInsider (June 2026), the new engine is specifically designed to run on-device using Apple's Neural Engine cores, ensuring that heavy computational tasks do not compromise system responsiveness. This focus on performance is reflected in other OS 27 features, where Apple claims photos will load up to 70% faster and local app performance has been optimized to improve overall efficiency. The removal of the manual color noise reduction slider in RAW 9 further highlights a move toward automated, AI-managed workflows where the system handles chroma noise by default. While Apple's integration simplifies the workflow, it enters a competitive landscape where Adobe and Topaz Labs have also deepened their AI offerings. Per Adobe (late 2025) and early 2026 reporting, Photoshop has recently integrated Topaz Labs' Sharpen and Denoise models directly as partner tools, though these often require generative credits or specialized subscriptions. Furthermore, independent testing of Adobe’s 2025 AI Denoise in Camera Raw has seen mixed reviews regarding detail retention vs. 'plastic' skin textures. Apple’s decision to perform denoising during the raw reconstruction phase is a technical strategy previously championed by DxO's DeepPrime technology, suggesting a direct challenge to premium third-party RAW converters that target professional photographers and cinematographers.
Read full article at redsharknews.com